AI Law - International Review of Artificial Intelligence Law
G. Giappichelli Editore

06/05/2024 - Generative AI and Data Scraping: guidelines from the Dutch Data Protection Authority

argument: Normativa/Regulations - Personal Data Protection Law


The Dutch Data Protection Authority has issued guidelines impacting companies that use data scraping to train generative AI systems. The guidelines assert that the only viable legal basis for data scraping might be "legitimate interest," even if the data is public and quickly deleted after use. The authority challenges the compatibility of data scraping with subsequent data processing and highlights the necessity of considering ethical implications and managing data responsibly. Special focus is given to the ethical need for transparency and the handling of special categories of data, stressing that commercial purposes alone do not justify data scraping unless they align with broader benefits to data subjects.